Second Homework: Cosponsorship
Seminar The Swiss Parliament in Comparison
Task 1: Cosponsorship networks
Figures 1 and 2 display the network graphs for the national council in the 48th (25% cosponsorship threshold) and 50th (33% cosponsorship threshold) legislature. Two marked differences between the two are that the second shows far fewer within-party connections for all major parties (except SP) and far fewer cross-party connections in the group Mitte-FDP-SVP.
The second graph suggests that SP has the most connections and Mitte the least (relative to the parties’ numbers of MPs). In fact, Table 1 shows that SP MPs have the most connections and significantly more than FDP MPs, who have a similar amount of connections as Mitte and SVP, as can be seen in Figure 3. On average, SVP MPs have the fewest connections, though not significantly fewer than FDP.
| Table 1: Effect of party on MPs' degree count (baseline: FDP) | |
|---|---|
| (1) | |
| (Intercept) | 6.89 *** (1.38) |
| GLP | 4.83 (3.42) |
| Grüne | 3.18 (2.60) |
| Mitte | 0.61 (1.90) |
| Other | 5.11 (2.85) |
| SP | 9.62 *** (1.79) |
| SVP | -2.08 (1.71) |
| N | 228 |
| R2 | 0.23 |
| *** p < 0.001; ** p < 0.01; * p < 0.05. | |
Task 2: Cross-party cosponsorship
I define an MP’s within-party support as the number of cosponsorships they have recieved of MP’s of their party, divided by the number of bills they have introduced and by the number of MPs in their party.
I define an MP’s cross-party support as the number of cosponsorships they have recieved of MP’s of other parties, divided by the number of bills they have introduced and by the number of MPs in other parties.
For this, all eleven MPs of minor parties are lumped together and I will not interpret their results.
Figure 4 shows all MPs’ within and cross-party support. There is a large cluster with small values on both axes and a handful of outliers with relatively high within and/or cross-party support. Christa Markwalder (FDP), former president of the national council, has the highest cross-party support (and also the third highest within-party support). Her 6 bills received 95 cosponsorships by MPs of other parties. This high value is mainly driven by one of her bills (on individual taxation of married couples) receiving 103 overall cosponsorships, more than half the national council.
Figures 5 and 6 below show the predicted values of within and cross-party support for each party. All major parties have similar levels of within-party support, while there is a clearer order for cross-party support: FDP, followed by Mitte, followed by SVP, GLP and SP, followed by Grüne. In fact, FDP’s support is significantly higher than that of SVP, SP and Grüne (all p-values around 0.02).
Task 3: IRT on cosponsorship data
The IRT places all MPs on a scale based on which bills they did and didn’t cosponsor. Figures 7 shows the MPs’ cosponsorship scores. A striking finding is that the biggest distance is between Mitte and SP/Grüne, with SVP/FDP lying in between (but much closer to Mitte). The network plot in Figure 1 suggests that the least cosponsorship is between SVP and SP/Grüne. I interpret the IRT’s placement of Mitte/SVP/FDP as follows: Since the parties’ degree of within-party cosponsorship is more similar than their cross-party cosponsorship (see Figures 5 and 6), the difference in the IRT should be driven more strongly by differences in cross-party cosponsorship. Mitte often cosponsors FDP and SVP, who’s bills are more right-wing than its own. Mitte’s less right-wing bills in turn get cosponsored by FDP and SVP, which is why the IRT places them closer to the center.
The one-dimensionality lumps within and cross-party cosponsorship together, a second dimension would probably help to untangle the two.